Wealthy democracies don’t become dictatorships. For a generation that adage has provided one of the firmest laws of modern democratization, the equivalent for comparativists of the democratic peace among international relations scholars. Like any big claim, the link between economic wealth and democratic durability has provoked debate. Political scientists have parsed the data, questioned the mechanisms involved and pursued new projects that validate the proposition even as they refine it. They have explored whether wealth not only sustains democracy but also produces it, and whether the distribution and forms of assets matter more than their raw amount.

Even while students build new and more nuanced links between development and democracy, they have not severed the connection established by political scientists Adam Przeworski and Fernando Limongi in their agenda-setting 1997 article “Modernization: Theories and Facts.” Building on seminal work by Seymour Martin Lipset, the authors determined that countries at all levels of material prosperity had become democratic (i.e., shifted from dictatorship to democracy). However, countries had not become authoritarian (shifted from democracy to dictatorship) when their non-oil GDP per capita exceeded $8,043 (in 2005 constant prices adjusted for purchasing power parity).

Why $8,043? That was the level of development Argentina had reached in 1975 before a junta toppled President Isabel Péron (40 years ago this week). Her government ranks as the most affluent democracy to fall in Przeworski and Limongi’s study. Democracies with higher GDP per capita were self-sustaining.

Przeworski and Limongi’s data stopped in 1990, but their principal finding has held for a quarter-century. Democratic experiments have fizzled in Africa (as in Mali after the 2012 coup), in post-communist states (witness Russia under Vladimir Putin) — and even in Latin America (thanks to pseudodemocrats like Nicaragua’s Daniel Ortega). But non-oil wealth in these circumstances was below the Argentina 1975 level.

Meanwhile, medium-income and richer democracies have remained dictator-proof — until now. The example of Turkey under premier-then-president Recep Tayyip Erdogan presents a potentially theory-busting specimen of a highly developed democracy going authoritarian. Despite recent market uncertainty, Turkey is now significantly more affluent than Argentina was 40 years ago and its political trajectory carries global implications. The more Erdogan and his Justice and Development Party (AKP) pull their economically vibrant country toward autocracy, the bleaker the outlook for democracy in similar or less favorable circumstances.

To be clear, Turkey’s democratization has never been seamless and Erdogan has not yet become a full-blown despot like Putin. After the ruling Republican People’s Party lost elections and peacefully passed power to the opposition in 1950, Turkey joined the political science canon as a multi-party Muslim-majority state — with periodic military interventions. In 1960, 1971 and 1980, the army pushed aside insufficiently secular governments, and then returned to the barracks. These interregnums disrupted democracy but conformed to Przeworski and Limongi’s pattern. At the time of the three coups, GDP per capita was approximately $3,200, $4,500, and $5,300, still significantly below the threshold at which development seals democracy.

Erdogan’s current drive, toward a super-powered presidency, is subtler than Turkey’s prior authoritarian periods. Rather than cancelling elections, Erdogan has employed the ballot box to quell his critics. Further, if electoral authoritarianism means only opposition parties lose on Election Day, then Turkey skirts the line. Last June, the electorate denied AKP parliamentarians the majority they had enjoyed since 2002. When neither the AKP nor any other party had formed a government by August, Erdogan called snap elections for that November.

Meanwhile state security forces aggressively battled the Kurdistan Workers’ Party (PKK) and menaced the AKP’s electoral rival, the Peoples’ Democratic Party (HDP). As fear spiked, media coverage and campaigning withered. A climate of insecurity probably helped the AKP rebound, winning half the votes and 58 percent of the seats. Nonetheless, Freedom House considered Turkey an electoral democracy in 2015. If the state keeps attacking judges, journalists and professors between voting cycles, Turkey may soon lose that designation (not that the country’s president would apparently mind).

If Erdogan ultimately confirms analyses that he is building a “competitive authoritarian” regime, his actions will be less novel than the context in which he acted. The country’s generals (briefly) led Turkey when it was relatively poor (and hence more prone to authoritarian reversals). Erdogan, though, has operated in socioeconomic conditions that should keep power dispersed. In 2003, when Erdogan became prime minister, GDP per capita reached $8,300. Since that time the economy has grown rapidly, adding more than six million jobs in recent years while reducing the “extreme poverty rate” from 13 percent to 4.5 percent.

Turkey’s middle class — historically a bulwark of democracy — has doubled in size. As Turks became better off, they ought to have become more effective at holding their leaders accountable. Instead, the opposite has occurred: Erdogan, buoyed by a cross-class alliance between the poor and the bourgeoisie, has combined economic advancement with political regression.

Authoritarianism in Turkey would not only upend Przeworski and Limongi’s findings, it would cast a pall over other high-income countries where democracy seems assured. A recent series of coups and autogolpes has raised concerns about an “authoritarian resurgence.” Yet most of the incidents in question, while troubling, are not iconoclastic. Among the countries Stanford University professor Larry Diamond listed in a 2015 essay on the topic, only one (Turkey) had reached the level of non-oil wealth per capita of Argentina in 1975 (see Table). Few people are happy when democracy teeters in locales as varied as Honduras, Niger and Sri Lanka. Nonetheless, anti-democratic currents are common in such developing economies. By contrast, Turkey in 2010 (the latest year of comparable data) belonged to a significantly more industrialized cohort, with a GDP per capita more than a quarter greater than Argentina’s in 1975. (Turkey was more affluent still in 2014, the year Diamond dates the breakdown of its democracy.)

Skeptics that development shapes regime type may see these figures as one more reason to reject the theory outright. For students who do believe wealth has historically preserved democracy, Turkey turning autocratic would shift expectations. Specifically, it would signal that in democracies with greater than $8,043 GDP per capita, the “hazard rate” (the probability that a democracy will become authoritarian in a given year) has risen above zero. Hence, current democracies with GDP per capita similar to Turkey’s—for example Bulgaria ($10,600), Romania ($9,400), or Brazil ($8,100) — would not be out of the woods. Authoritarianism returning in such developed places has seemed impossible, but Erdogan’s tenure suggests otherwise. Further, if Turkey’s experience invites vigilance in medium-to-high income states, then it implies added concern for significantly less developed democracies — like Albania ($6,600), Tunisia ($6,100), and Paraguay ($4,100) — that have not even reached the level of Argentina 1975.

It is too soon to know whether Turkey under Erdogan will be an outlier or a bellwether. Already, however, Turkey’s experience suggests that the economic forces that previously bolstered democracy appear to be weakening, perhaps dramatically. The causes of this shift — and whether it can be offset by stronger opposition parties and tighter constraints on executive power — remain to be determined.

Jason Brownlee is professor of government at the University of Texas at Austin and the author of three books, including (with Tarek Masoud and Andrew Reynolds), “The Arab Spring: Pathways of Repression and Reform” (Oxford University Press, 2015). His current project is titled Democratization in an Age of Authoritarianism.